Hello Lee,

Friday, May 03, 2002, 4:21:04 PM, you wrote:

DL> I'm not on the dev team or anything, but I don't understand how you could have 
*ever* successfully encrypted multiple streams with the same EVP context??? Just the 
IV's alone would have been
DL> screwed up for CBC ciphers and stream ciphers like RC4 would completely break.

ECB won't - IV isn't needed there... I didn't say its a complete solution. Still it 
doesnt
break nothing as far as I can tell. More changes has to come of
course... Still what exactly this patch does is:

I EVP_UpdateEncrypt is called EVP_EncryptFinal intermidiate data in
the temporary buf in the context wont be considered a part of the new
data stream.

DL> Unless, of course, you decrypted in *exactly* the same sequence of streams as you 
encrypted, in which case, each individual stream would be incorrectly encrypted.


DL> still, I don't see how this could have ever worked.

DL> -lee

DL> -----Original Message-----
DL> From: Pavel Tsekov [mailto:[EMAIL PROTECTED]]
DL> Sent: Friday, May 03, 2002 5:27 AM
DL> To: [EMAIL PROTECTED]
DL> Subject: Re: [PATCH] Re: Question on EVP encryption/decryption routines


DL> Is there any chance that somone will comment on this topic anytime
DL> soon ?

PT>> Does the attached oneliner make sense ? IMO it doesnt break anything,
PT>> but allows reusing the EVP_CIPHER_CTX in calls to EVP_EncryptUpdate,
PT>> after EVP_EncryptFinal is called.

PT>> Any comments greatly appreciated :)

PT>> The diff is against the released 0.9.6c version of the OpenSSL
PT>> library.

PT>>> Currently EVP_Encrypt* interface doesnt support
PT>>> encryption of multiple data streams with a single
PT>>> symmetric key. Is there any chance patches to
PT>>> allow this to be accepted in the official OpenSSL
PT>>> codebase ? Currently I'm thinking to modify EVP_EncryptFinal
PT>>> so it leaves the EVP_CIPHER_STATE in a consistent state
PT>>> after finishing its work.

PT>>> Any comments on this ? I am aware of the fact that encrypting
PT>>> multiple data streams with a single key is not such a good idea
PT>>> of a security point of view.

DL> ______________________________________________________________________
DL> OpenSSL Project                                 http://www.openssl.org
DL> Development Mailing List                       [EMAIL PROTECTED]
DL> Automated List Manager                           [EMAIL PROTECTED]
DL> ______________________________________________________________________
DL> OpenSSL Project                                 http://www.openssl.org
DL> Development Mailing List                       [EMAIL PROTECTED]
DL> Automated List Manager                           [EMAIL PROTECTED]



-- 
Best regards,
 Pavel                            mailto:[EMAIL PROTECTED]

______________________________________________________________________
OpenSSL Project                                 http://www.openssl.org
Development Mailing List                       [EMAIL PROTECTED]
Automated List Manager                           [EMAIL PROTECTED]

Reply via email to